Background: Combining various aspects of child feeding into an age-specific summary index provides a first answer to the question of how best to deal with recommended feeding practices in the context of HIV pandemic. The objective of this study is to assess feeding practices of HIV exposed infants using summary index and its association with nutritional status in Southern Ethiopia. Methods: Facility based cross-sectional study design with cluster random sampling technique was conducted in Sidama Zone, Southern Ethiopia. Bivariate and multivariable linear regression analyses were performed to assess the association between summary index (infant and child feeding index) (CS-ICFI) and nutritional status. Results: The mean (±standard deviation (SD)) cross-sectional infant and child feeding index (CS-ICFI) score of infants was 9.09 (±2.59), [95% CI: 8.69-9.49]). Thirty seven percent (36.6%) of HIV exposed infants fell in the high CS-ICFI category while 31.4% of them were found in poor feeding index tertile. About forty two percent (41.6%) of urban infants were found in the high index tertile but only 24% of the rural infants were found in high index tertile. Forty six percent (46%) of the rural infants were found in low (poor) feeding index category. The CS-ICFI has a statistically significant association with weight for age z score (WAZ) (ß = 0.168, p = 0.027) and length for age z score (LAZ) (ß = 0.183 p = 0.036). However CS-ICFI was not significantly associated with weight for height z score (WLZ) (p = 0.386). Conclusion: Majority of HIV exposed infants had no optimum complementary feeding practices according to cross-sectional infant and child feeding index. CS-ICFI was statistically associated especially with chronic indicators of nutritional status (LAZ and WAZ). More rural infants were found in poor index tertile than urban infants. This may suggest that rural infants need more attention than urban infants while designing and implementing complementary feeding interventions.